NEW INC. MAGAZINE COLUMN FROM HOWARD TULLMAN - AND A.I. ENGINE DEMOS

 

How to Start Smart With a Plug and Play AI Model

There is a better approach to AI for businesses that don’t want to shell out millions of dollars to Google or OpenAI to build their own custom version of ChatGPT.

EXPERT OPINION BY HOWARD TULLMAN, GENERAL MANAGING PARTNER, G2T3V AND CHICAGO HIGH TECH INVESTORS @HOWARDTULLMAN1

Oct 7, 2025

It’s hard to say whether there are more columns arguing that, if you haven’t already infused your business processes with AI you may be too late and consequently doomed, or that the vast majority of corporate AI projects to date have been miserable and costly failures. The latter parade of horribles is led right now by recent M.I.T pronouncements asserting that something like 95 percent of the serious large organization efforts have failed.  

There are plenty of reasons offered for the failures of these initiatives, but they sound unsurprisingly similar to the explanations offered when the pace of adoption and implementation of any new tool or technology is slower than the hyped expectations—and when the promised payoffs in savings and enhanced productivity among others fail to materialize immediately. The real wonder is why resistance to change comes as any surprise to anyone at this late date.  

In the case of AI, there is another debilitating and discouraging aspect of the process, which is that even after all the time and analysis and reviews aimed at attempting to determine why some given project hasn’t worked, there’s almost nothing to be gathered or learned because the “black box” that you’re attempting to interrogate and explore is just that—a black box which performs its function in somewhat mysterious and unclear ways that no honest auditor would even pretend to fully understand. 

Needless to say, it’s hard to extract lessons or process improvements in order to achieve better future results when the underlying engine driving the outputs (a) has no memory, (b) doesn’t really “learn,” and (c) doesn’t accept or incorporate input, instructions, and corrections in any consistent or measurable fashion. It just is what it is and you’re offered the opportunity to take it or leave it. This isn’t the most appealing context or foundation to try to build the next phase of your business upon.  

To be clear, the opaqueness and obstinance of the primary AI foundational platforms aren’t even the most pressing concerns for most smaller companies and businesses which are trying to incorporate AI into their organizations. These firms — which is to say the vast majority of all the entities in the world except the mega-corporations — can’t afford to spend their scarce and critical resources on broad AI experiments which may or may not ever come to fruition. Even more importantly, they are facing a much more pressing set of risks and concerns within their own companies presented by the actions of their own people.  

Close to 60 percent of employees are using AI tools not approved by their employers, according to a recent Cybernews survey. Worse yet, more than half admit that their direct supervisors or managers know about their use and don’t object, while another sizeable group reports that their managers don’t care. In fact, the survey found that it was executives and senior managers who were the most likely to be using these tools. But the critical finding from the survey is that around 75 percent of the respondents using unauthorized tools admitted to sharing sensitive data with them. And the companies themselves have no control or even information about which team members are doing what. Once the data enters any of these AI platforms, the user has no control over whether the information is stored, reused or exposed to third parties. 

There is a much better and smarter approach for SMBs and, frankly, for just about any business not willing to shell out millions of dollars to Google or tens of millions of dollars to OpenAI to build their own custom version of ChatGPT. That’s what I call the “plug and play” solution. Simply stated, this is a “portable” query engine that any business can install on their own premises and in their own machines and networks.  

Typically, there is an initial fee for installation and implementation of the system—roughly $50,000 or less—and then a monthly or annual recurring cost which may vary with the volume of usage but which isn’t likely to be a major cost and can be cancelled at any time. This puts a known cap and realistic limitation on the entire cost of the experiment. 

The firm designates and assigns an underlying selection of data for the engine (the corpus) to operate on, interrogate, and extract answers from. All employees have access to the system and can simply ask whatever questions about the data which they need further information about, request analytics or compilations of segments or selections of the data, or request the creation of documents, reports, or presentations of parts of the data which are relevant to areas or questions they are dealing with. One key consideration is that by limiting the dataset to material relevant to the specific business and activities of each company, the fit of the engine to the likely inquiry scope is better and the overall costs of operation as well as response time are reduced since there is no need to “search worldwide” or attempt to “boil the ocean” in order to develop and respond with timely and accurate information.  

Note here that the system has two critically important guardrails in place which govern every query. First, it knows what it knows, so it’s designed to report back that it can’t answer certain questions. Second, it knows the scope and limitations of what queries are proper topics for inquiry and rejects irrelevant or immaterial prompts. For example, the system doesn’t know the meaning of life. In addition, the system creates a log file of every query and every answer which can be reviewed by management to track activity and also readily determine how often and how effectively the system is being used.  

Other crucial elements of these systems are that there are evaluation, review and edit functions which in essence let the system “learn” and improve its answers and responses on an ongoing basis as well as allowing for corrections and updates—all within the control of the company itself and without ever allowing any such proprietary information to leave the premises. Security and control as well as demonstrable growth in efficacy are all central to the system. Because the system also permits any user to evaluate and rank the value and accuracy of any answer, there is a real time feedback loop that encourages and rewards employee use and engagement. This prospect makes it far more likely that team members will adopt and regularly employ the system especially as they see their own input and roles in the process incorporated into the system.  

Bottom line: you can put a toe in the water at a reasonable cost and try to expose your people to the upsides of these next-generation tools without incurring substantial expenses, without exposing your proprietary data to the outside world in an uncontrolled and unaccountable manner, and without encouraging or permitting your own team members to go outside the walls and waste time and energy randomly exploring the large LLMs without any particular guidance, support or benefit to your firm.  

I have installed two sample systems on my own website in the lower right-hand corner. These are free to use and to experiment with for all readers. The white link has a corpus which consists solely of my hundreds of Inc. magazine columns published over the last decade. The black link is a much broader corpus of all my books, columns, speeches and presentations, etc. which will attempt to answer a far broader range of questions. Give it a try before you decide to buy. Test before you invest.

 

New INC. Magazine column from Howard Tullman 4 Reasons You Need to Right-Size Your AI

 

4 Reasons You Need to Right-Size Your AI

Sometimes it’s smart to be careful—and slow—when you’re dealing with new technologies.

EXPERT OPINION BY HOWARD TULLMAN, GENERAL MANAGING PARTNER, G2T3V AND CHICAGO HIGH TECH INVESTORS @HOWARDTULLMAN1

Aug 26, 2025

This is a very complex and challenging time for startups and small businesses in terms of how they should be addressing all of the issues and concerns around artificial intelligence and, more specifically, how they can incorporate the new AI tools and technologies into their own businesses. I realize that every startup in the world already professes to have built AI into their current offerings as well as into their future plans but, at best, many of these claims are nothing more than adaptations of machine learning or pattern recognition with a new shiny coat of paint and some text prediction capability. Sometimes it’s smart to be careful and slow when you’re dealing with new technologies. The last thing you want to do is be the latest victim of the fake-it-‘til-you-make-it disease.

It’s not remotely clear that a surface-level solution built on top of a generic large language model system will be of much value or benefit to many midsize businesses with very specific needs and nuanced market dynamics. One size almost never fits all these days. The implementation and operating costs alone of many of these systems would likely exceed any concrete internal improvements that addressed the user’s real needs. On the other hand, a smaller, more targeted, and clearly focused system whose objectives and functions the company’s management understands could be a valuable aid and time-saver if properly deployed.  

A side note that should be obvious but is often overlooked in top-down implementations of new tech is that you must secure buy-in from your key management and other pivotal team members and address in advance their concerns and the typical misunderstandings they may have about the plans, the short- and long-term job consequences, and other implications of the new systems and their roles in the process. 

We’re all rushing to employ these things before we fully understand them and, worse yet, it’s easy to come to depend on these seductive tools even when we know in our hearts that we’re not fully in control of them. You don’t need to cross the chasm in a single bound. Hallucinations and biases are only two of the most obvious risks and concerns when you start looking under the hood of some of these programs and discover that even their makers have only a passing idea of how they really work. 

The big guys in the corporate world can now rush to join the line of lemmings willing to pay OpenAI a consulting fee starting at $10 million to send a team of its eager engineers into their shops to build them custom solutions based on its GPT-4o technology. You would think that—given the havoc that the DOGE monkeys and minions brought about across our entire government—these corporate honchos would take a breath or two and ask themselves whether turning over the keys to their futures to Sammie’s smarties is the wisest course or whether it’s roughly akin to giving expensive whiskey and your car keys to the neighbor’s teenage son and wishing him well on his journey. 

If there’s a single statement that says it all for me right now, it’s the various versions of the observation that no one’s going to lose their business to AI, but most will lose their businesses to competitors who are more effectively using AI to streamline and accelerate their operations, to reduce their headcount without sacrificing customer connections and satisfaction, and to give them a far broader and more accurate overview of their marketplace, their competition, and timely intelligence and data to react to emerging positive and negative trends. 

The best and quickest of the players will rapidly realize that the hours and days they previously spent pouring over voluminous market data, analyzing their often incomplete and delayed compilations, and attempting to extract actionable information from the mess will now be replaced and made available in real-time detailed summaries crafted by young and clever prompt engineers.

The truth is that—with regard to the introduction of any new and disruptive technology—it will take every business a significant amount of time to learn how best to deploy it and how to deal with the displacements, interruptions, and new responsibilities and job descriptions that will accompany it and inevitably cause problems.  

Walking before you run—especially if you’re trying to do this development and implementation basically on your own—is the only rational and cost-effective course. It’s critical to keep in mind that you can always circle back and build better and more robust versions of what you’re initially experimenting with. It’s not likely to be an overnight project or an overnight success, but each iterative step will teach you a great deal, further empower you, and also help you to better understand the capabilities of the tools you are using—even as those abilities continue to grow and expand every day.  

What’s most important is for you to take the time to gather your team and review your operations and outline the areas where some intelligent automation could speed and simplify your own processes and actually produce a better result. In the first instance, none of this needs to be rocket science. Guesty is a legitimately AI-assisted property management system that was designed specifically for short-term rentals handled by Airbnb owners and operators.  

While this sounds about as mom-and-pop as can be, these folks face many of the same issues you do in your businesses—albeit at perhaps a smaller scale. The point is that, if this kind of simple use-case can show dramatic improvements in their metrics and their bottom lines, then shame on you if you haven’t figured out how to replicate these tools and techniques in your own shop.  

Here are four simple examples that a satisfied Airbnb operator told me has increased his yield and profit, dramatically decreased the time he was spending each week on his side business, improved his ratings and rankings with Airbnb, and led to repeat business and referrals from satisfied customers. And to be clear, I think he spends about $30 a month for the app. Eat your heart out.  

1. Hundreds of stored FAQ responses are delivered automatically in context-sensitive and narrative serial fashion 

You would be surprised and possibly shocked to learn how many times a day your team members waste their time repeatedly responding to and answering the same questions over and over again. Often, they do it slowly or inaccurately and eventually they do it impatiently—human nature being what it is—and none of this is good for your business. Automated responses can satisfy a significant number of callers who have simple, redundant inquiries and, more importantly, can deflect the wrong callers by simply and quickly making it clear to them that they are looking in the wrong place. 

Pricing is dynamic 24/7 and throughout each week based on a variety of factors and competitive offerings in the market as well as available capacity 

While in theory you could spend your entire day checking out competitive offerings and prices and adjusting your offers accordingly (and clearly Amazon does its pricing in this fashion every minute) and you could also constantly check your bookings through the week and determine whether price reductions might absorb available and empty units, rooms, or beds (just as American Airlines does all day long), the fact is that neither you nor anyone on your team has the time or interest to do anything like this, but the Guesty system does it automatically for you according to your guidelines and parameters instantly every day.  

Publish and sync your listings in real time across more than 50 major listing services including all the major sites  

You may use programmatic tools (with very little actual accountability) to get your messages out to the masses, but, in truth, you have little idea of who is seeing them and absolutely no real time ability to change or update the content or distribution plan. Intelligent systems using open APIs across multiple platforms give you a one-stop solution to precisely target and deliver your messages to qualified, interested viewers in the proper context with the ability to vary and alter any portion of the listings that you wish at any time.  

Responses to every inquiry are immediately replied to even if the reply is merely a placeholder and conversation starter 

Not surprisingly, response time is a measurable metric that firms like Airbnb use to evaluate the owners and operators on their site who use their services. Automated intelligent systems can respond instantly to every inquiry even if the response isn’t a substantive answer, but only a request for further info, details, or specificity to continue the conversation. In addition to managing the Airbnb metric, this immediate reply improves customer satisfaction and engagement without consuming any incremental resources until the lead is further qualified.  

Bottom line: While these examples may not directly apply to your company’s needs and current operations, each of them is an invitation and a suggestion to explore similar kinds of concerns and friction within your own organization and to see how AI and intelligent automation can help to address and improve things.  

NEW INC. MAGAZINE COLUMN ON A.I. FROM HOWARD TULLMAN

 

Artificial Intelligence Embeds Across Sectors Without Heed to Its Human Impact 

AI is becoming a part of of our everyday lives, whether we like it or not. 

EXPERT OPINION BY HOWARD TULLMAN, GENERAL MANAGING PARTNER, G2T3V AND CHICAGO HIGH TECH INVESTORS @HOWARDTULLMAN1

JAN 28, 2025

Senior education officials, regulators, and media mavens all over the world have been focused for some time on the issue of how teachers will be able to distinguish between materials written by students and those created by technologies driven by artificial intelligence. Interestingly, the majority of educators who work in the field every day with students don’t think this is much of a concern. They know their students, they know their respective abilities and capacity, and frankly they only wish their students were smart, motivated and talented enough to try to accomplish such a feat of prevarication.  

Another interesting discussion is taking place in the work world. It’s everywhere. Consider the controversy from the movies, which are now all a twitter (no pun intended) about the AI-based voice enhancement technology used to improve the authenticity of the Hungarian voices in The Brutalist movie. Then there’s the magazine world: the not-too-distant but humiliating discovery in 2023 that articles in Sports Illustrated were actually AI-written and attributed to non-existent authors. Which, by the way, also had headshots. No one complained about the content of the stories, they were just apparently horrified by the process of computers replacing copywriters.

All these concerns stem from fears arising in two different areas. First, there is anxiety across industries about job elimination through automation and A.I. implementation. And, second, the increasingly prevalent idea that we are all less able to tell the difference in so many ways between men and machines.

Plenty has been written about job losses, but we’re just beginning to realize how exposed and how unaware we are of the extent to which our expanding and encroaching technologies have subtly and unobtrusively invaded and subsumed so many aspects of our day-to-day lives. One of the most simple and obvious examples is captchas. We now take for granted and unironically that it’s become our daily job to repeatedly prove to computers that we are real human beings before they permit us to get on with so many different activities and transactions. For the moment, it seems that we’re all stuck with technology, when all we really want is stuff that works.

Real-World Insight

The problem is that our technology development work is so completely focused on the future that we seldom, if ever, look backward. As a result, rather than learning from mistakes, we are doomed to keep repeating them and forgetting the lessons that we should have painfully learned by now. As a result, we quickly come to depend on these new modes of assistance and support. At the same time, we become fearful because we know that there are aspects of their operation and abilities that we can’t entirely control. I’m not talking about Skynet and Arnold. But some more subversive undertakings are superficially attractive, clearly less threatening at the moment. These are designed to replicate, impersonate, and deal directly with other machines and computers “as if” they were human.

With the announced and accelerating rollouts of agentic tech, I believe that we’re on the cusp of another deep technology rabbit hole which we’re largely unprepared for and ill-equipped to deal successfully with. What we never seem to appreciate is that when we develop new disruptive tools and technologies, we immediately seize on the initial implementations and put them into action before we remotely understand them in their entirety. Much less consider their unforeseen and consequential longer-term effects, or even appreciate how long and costly a process will be required to understand how to best put them to use. Every new technology is a package deal, which brings its own negativity right along with all its benefits.

The recent unveiling by OpenAI of its new agentic offering called Operator is the latest clear step forward, for better or for worse. Incorporating computer-using agency, along with the ability to interpret and act upon handwritten lists and other images, Operator – for all intents and purposes – looks to other computers like a human operator who is using both a keyboard and a mouse. Already connected to Open Table and Instacart among other apps and services, Operator can seamlessly book tables and reservations, order tickets, select groceries, and initiate regularly scheduled tasks with very limited, if any, human intervention once the process is set in motion. Only at the final moments and specifically when payment information and confirmation is required does the system pause and ask for approval before proceeding. It’s only a short further step to complete autonomy and reaching the point where, as the late great singer-songwriter Jim Croce sang in his version of his hit Operator, “There’s no one there I really wanted to talk to.”

AI Anxiety

If this prospect doesn’t recall the frightening scenes from Fantasia where the unstoppable brooms carrying buckets of water marched ceaselessly forward and step right over poor Mickey, the Sorcerer’s Apprentice, then you’re simply not old enough or a fan of classic Disney movies. Embedded in this fantasy is a real warning which has even more direct and important application today. It’s not difficult to imagine even more sophisticated and fully automated onslaughts launched against ticket sellers or new and more convincing scams and frauds using data and imagery extracted by these new tools.

A photo of a handwritten shopping list – as used in the Operator demo video – seems innocent and harmless until you realize that you’ve provided the digital world with the ability to readily replicate your cursive signature. This may matter less as we move forward, and the schools completely abandon any effort to teach our kids how to sign their names on documents or even write properly and settle instead for block printing.

Bottom line: Here we go again on a wild chase into the future without any clear end in sight or a sufficient understanding of the risks involved or how they might be limited or circumscribed. We’re buying the ticket, closing our eyes, and taking the ride. As the late Hunter Thompson used to say: “There is no honest way to describe the edge because the only people who really know where it is are the ones who have gone over.”